[SaadShakeel1]

Summary

Completes TaskFlow’s core backend and adds AI-powered project risk analysis using
Groq.

What changed

• Added persistent AI analysis runs with queued, running, succeeded, and failed
  states.
• Added project snapshots containing authorized project, task, dependency, and
  activity data.
• Added a background worker that safely claims queued jobs with PostgreSQL SKIP LOCKED.
• Added Groq structured-output analysis with validated findings and
  recommendations.
• Added AI analysis APIs:
  • POST /projects/{id}/analysis-runs
  • GET /projects/{id}/analysis-runs
  • GET /projects/{id}/analysis-runs/{runID}
• Added model, API, worker, Groq-client, and opt-in live integration tests.
• Added public repository documentation, .env.example, GitHub Actions CI, and
  MIT license.

Security

• API keys remain local in .env and are ignored by Git.
• AI receives only a scoped project snapshot—never passwords, sessions,
  verification tokens, or database credentials.
• AI findings are advisory and cannot mutate project data automatically.

Verification

• go test ./... -count=1
• go vet ./...
• git diff --check
• Live Groq structured-output integration test passed using a synthetic project
  snapshot.